The reviewed record of science sign in
Pith

arxiv: 2606.01945 · v1 · pith:SHMLKVVL · submitted 2026-06-01 · cs.CV

Beyond Low-Rank: Low-Rank Sparse Prompting via Spiking Neural Network and Prompt Factorization

Yumiao Zhao , Bo Jiang , Beibei Wang , Xixi Wan , Xiao Wang , Jin Tang This is my paper

Reviewed by Pith T0 review T1 audit T2 compute T3 formal T4 kernel 2026-06-28 15:15 UTCgrok-4.3pith:SHMLKVVLrecord.jsonopen to challenge →

classification cs.CV
keywords visual promptingspiking neural networkslow-rank factorizationsparse promptsmodel adaptationcomputer visionefficient fine-tuning
0
0 comments X

The pith

Combining low-rank factorization with spiking neural networks generates instance-specific sparse visual prompts for more compact model adaptation.

A machine-rendered reading of the paper's core claim, the machinery that carries it, and where it could break.

The paper aims to establish that visual prompts can be made sparse and instance-specific by processing low-rank prompt factors through spiking neurons that fire only when thresholds are met. This approach maintains the benefits of low-rank structure while adding sparsity from the discrete spike trains produced by the integrate-and-fire mechanism. A reader would care because current visual prompting often uses dense prompts that add redundant changes and require many parameters to learn. If the method works, it would allow pre-trained vision models to adapt to new tasks using fewer adjustable parameters and with better generalization across different benchmarks.

Core claim

LoRSP constructs prompt factors via low-rank factorization and feeds them into an SNN that performs integrate-and-fire to emit spikes, thereby generating a sparse visual prompt while keeping the low-rank constraint. This enables instance-specific selective prompting that leads to more compact and robust adaptation across diverse downstream tasks, as shown in experiments on five vision backbones.

What carries the argument

The spiking neural network's integrate-and-fire dynamics applied to low-rank prompt factors, which converts continuous inputs into discrete sparse spike trains for prompt generation.

If this is right

  • LoRSP achieves competitive performance on multiple benchmarks while using fewer tunable parameters than existing visual prompting methods.
  • The sparse prompts are instance-specific due to the selective firing of spiking neurons.
  • Adaptation becomes more compact and robust across heterogeneous vision backbones and tasks.
  • The low-rank constraint is preserved even as sparsity is introduced.

Where Pith is reading between the lines

These are editorial extensions of the paper, not claims the author makes directly.

  • If the sparsity improves generalization, similar spiking mechanisms could be tested on language model prompting to reduce parameter counts.
  • The energy efficiency from sparse spikes might make this suitable for edge device adaptations, though not tested in the paper.
  • Further work could examine whether the low-rank factors capture all necessary prompt subspaces or if higher ranks are sometimes needed.

Load-bearing premise

The integrate-and-fire process in the spiking neural network will reliably turn low-rank prompt factors into sparsity patterns that reduce tunable parameters and improve generalization over dense low-rank prompts.

What would settle it

If experiments on the same benchmarks and backbones show that LoRSP uses equal or more parameters or achieves lower accuracy than standard low-rank visual prompting methods, the advantage of the spiking sparsity would be called into question.

Figures

Figures reproduced from arXiv: 2606.01945 by Beibei Wang, Bo Jiang, Jin Tang, Xiao Wang, Xixi Wan, Yumiao Zhao.

Figure 1
Figure 1. Figure 1: Illustration of various Visual Prompting (VP) paradigms. (Left) [PITH_FULL_IMAGE:figures/full_fig_p001_1.png] view at source ↗
Figure 2
Figure 2. Figure 2: The overall framework of the proposed Low-Rank visual Spike Prompting (LoRSP). LoRSP first applies low-rank factorizations to generate group￾wise prompt factors. These factors are then fed into the SNN module to generate pixel-level sparse prompts for each instance. Spiking-PointNet [43] adopts SNNs for 3D point cloud pro￾cessing, leveraging the inherent sparsity of spiking neurons to significantly allevia… view at source ↗
Figure 4
Figure 4. Figure 4: Effects of the Vth and β in LoRSP on two backbones (ViT-B/16 and Swin-B). factors on both backbones. Increasing the number of prompt factors ( k > 4) or the rank (r > 4) leads to a slight drop in performance. These results indicate that a larger rank or a greater number of prompt factors can introduce redundant prompt components, while small k or r limit the capacity to capture instance-specific cues. Ther… view at source ↗
Figure 3
Figure 3. Figure 3: Impact of the number of prompt factors and rank in LoRSP and [PITH_FULL_IMAGE:figures/full_fig_p007_3.png] view at source ↗
Figure 5
Figure 5. Figure 5: Visualization of learned visual prompts. We compare LoR-VP and our LoRSP by visualizing the input images, the learned visual prompts, and the [PITH_FULL_IMAGE:figures/full_fig_p008_5.png] view at source ↗
read the original abstract

Visual Prompting (VP) has emerged as an efficient paradigm for adapting large-scale pre-trained vision models to downstream tasks by incorporating learnable prompts at the input level. However, existing VP methods typically employ dense pixel-level prompts, which often suffer from redundant perturbations, limited generalization and energy inefficiency. To overcome these limitations, we propose to integrate brain-inspired spiking learning into visual prompt learning tasks. As we know that spiking neuron can perform inexpensive information processing by transmitting the input data into discrete spike trains and return sparse outputs. Inspired by this, we propose \textbf{Lo}w-\textbf{R}ank visual \textbf{S}pike \textbf{P}rompting (LoRSP), a novel framework that learns dynamic low-rank sparse visual prompts naturally via a Spiking neuron learning mechanism. The core idea of LoRSP is to exploit the brain-inspired sparse firing mechanism of spiking neurons to generate pixel-level sparse prompt for each instance. To be specific, we first construct a series of prompt factors via low-rank factorization to capture distinct prompt subspaces. These prompt factors are then fed into an SNN architecture, which performs the integrate-and-fire process to emit spikes. As a result, our LoRSP generates a \emph{sparse} visual prompt while maintaining the low-rank constraint. This design enables instance-specific selective prompting, leading to more compact and robust adaptation across diverse downstream tasks. Extensive experiments on five heterogeneous vision backbones and multiple benchmarks demonstrate that LoRSP achieves competitive performance while requiring fewer tunable parameters compared to existing VP methods.

Editorial analysis

A structured set of objections, weighed in public.

Desk editor's note, referee report, simulated authors' rebuttal, and a circularity audit. Tearing a paper down is the easy half of reading it; the pith above is the substance, this is the friction.

Referee Report

2 major / 1 minor

Summary. The manuscript proposes LoRSP, a visual prompting method that factorizes prompts into low-rank factors, feeds them through an SNN with integrate-and-fire dynamics to induce sparsity, and claims to produce instance-specific sparse prompts that preserve the low-rank constraint, yielding competitive adaptation performance with fewer tunable parameters across vision backbones.

Significance. If the construction truly preserves low-rank structure after the nonlinear SNN step and the resulting sparsity yields measurable gains in generalization or parameter efficiency, the work would offer a principled way to combine low-rank efficiency with brain-inspired sparsity for visual prompting; however, the absence of any supporting derivation, equations, or results makes significance difficult to assess at present.

major comments (2)
  1. [Abstract] Abstract: the claim that the method 'generates a sparse visual prompt while maintaining the low-rank constraint' is load-bearing for the title and contribution yet receives no justification; the integrate-and-fire operation is a nonlinear threshold-and-reset map whose output, when used to reconstruct the prompt, has no obvious reason to remain low-rank.
  2. [Abstract] Abstract: the statement that 'extensive experiments on five heterogeneous vision backbones and multiple benchmarks demonstrate competitive performance while requiring fewer tunable parameters' is unsupported by any baselines, metrics, tables, error bars, or implementation details, rendering the performance claims unevaluable.
minor comments (1)
  1. [Abstract] Abstract: the expansion 'Low-Rank visual Spike Prompting' for LoRSP is inconsistent with the title phrasing 'Low-Rank Sparse Prompting'; standardize the acronym expansion.

Simulated Author's Rebuttal

2 responses · 0 unresolved

We thank the referee for the constructive comments on our submission. We address each major comment point-by-point below and commit to revisions that strengthen the justification and clarity of the claims.

read point-by-point responses

  1. Referee: [Abstract] Abstract: the claim that the method 'generates a sparse visual prompt while maintaining the low-rank constraint' is load-bearing for the title and contribution yet receives no justification; the integrate-and-fire operation is a nonlinear threshold-and-reset map whose output, when used to reconstruct the prompt, has no obvious reason to remain low-rank.

    Authors: We agree the abstract claim requires explicit justification, as the integrate-and-fire map is nonlinear. In the method, low-rank factorization produces the prompt factors first; the SNN then generates instance-specific spike trains that act as a sparse gating mechanism applied to those factors before reconstruction. This structure is intended to keep the effective prompt within the original low-rank subspace. However, the current manuscript does not include a formal derivation or equations demonstrating preservation of the low-rank property after the nonlinearity. We will add a dedicated subsection with the mathematical argument and any necessary assumptions in the revision. revision: yes

  2. Referee: [Abstract] Abstract: the statement that 'extensive experiments on five heterogeneous vision backbones and multiple benchmarks demonstrate competitive performance while requiring fewer tunable parameters' is unsupported by any baselines, metrics, tables, error bars, or implementation details, rendering the performance claims unevaluable.

    Authors: The abstract summarizes results presented in the experiments section, which reports comparisons against existing VP methods across the stated backbones and benchmarks, including parameter counts. That said, the referee is correct that the abstract itself provides no supporting numbers or references, making the claims difficult to evaluate from the abstract alone. We will revise the abstract to either tone down the language or include a concise reference to key quantitative outcomes (e.g., average accuracy and parameter reduction) while ensuring the main text already contains the full tables, baselines, and implementation details. revision: yes

Circularity Check

0 steps flagged

No significant circularity detected

full rationale

The paper introduces LoRSP as a novel combination of low-rank factorization on prompt factors followed by SNN integrate-and-fire for sparsity. The abstract and description present this as an architectural design choice without equations or claims that reduce the output prompt to a quantity defined by the inputs themselves. No self-citations, fitted predictions, or uniqueness theorems are invoked in a load-bearing way. The method is self-contained as a proposed framework rather than a derivation that collapses by construction.

Axiom & Free-Parameter Ledger

0 free parameters · 0 axioms · 0 invented entities

Abstract-only review provides no equations or implementation sections, so no specific free parameters, axioms, or invented entities can be extracted. The approach references standard low-rank factorization and SNN integrate-and-fire but does not detail any ad-hoc choices or new entities.

pith-pipeline@v0.9.1-grok · 5825 in / 973 out tokens · 29354 ms · 2026-06-28T15:15:52.664759+00:00 · methodology

discussion (0)

Sign in with ORCID, Apple, or X to comment. Anyone can read and Pith papers without signing in.

Reference graph

Works this paper leans on

63 extracted references · 10 canonical work pages · 1 internal anchor

  1. [1]

    An Image is Worth 16x16 Words: Transformers for Image Recognition at Scale

    A. Dosovitskiy, “An image is worth 16x16 words: Transformers for image recognition at scale,”arXiv preprint arXiv:2010.11929, 2020. JOURNAL OF LATEX CLASS FILES, VOL. 14, NO. 8, AUGUST 2021 9

    work page internal anchor Pith review Pith/arXiv arXiv 2010

  2. [2]

    Swin transformer: Hierarchical vision transformer using shifted windows,

    Z. Liu, Y . Lin, Y . Cao, H. Hu, Y . Wei, Z. Zhang, S. Lin, and B. Guo, “Swin transformer: Hierarchical vision transformer using shifted windows,” inProceedings of the IEEE/CVF international conference on computer vision, 2021, pp. 10 012–10 022

    2021

  3. [3]

    Clip-adapter: Better vision-language models with feature adapters,

    P. Gao, S. Geng, R. Zhang, T. Ma, R. Fang, Y . Zhang, H. Li, and Y . Qiao, “Clip-adapter: Better vision-language models with feature adapters,” International Journal of Computer Vision, vol. 132, no. 2, pp. 581–595, 2024

    2024

  4. [4]

    A closer look at the few-shot adaptation of large vision-language models,

    J. Silva-Rodriguez, S. Hajimiri, I. Ben Ayed, and J. Dolz, “A closer look at the few-shot adaptation of large vision-language models,” in Proceedings of the IEEE/CVF Conference on Computer Vision and Pattern Recognition, 2024, pp. 23 681–23 690

    2024

  5. [5]

    Prompt-and-transfer: Dynamic class-aware enhancement for few-shot segmentation,

    H. Bi, Y . Feng, W. Diao, P. Wang, Y . Mao, K. Fu, H. Wang, and X. Sun, “Prompt-and-transfer: Dynamic class-aware enhancement for few-shot segmentation,”IEEE Transactions on Pattern Analysis and Machine Intelligence, 2024

    2024

  6. [6]

    Relationship prompt learning is enough for open-vocabulary semantic segmentation,

    J. Li, Y . Lu, Y . Xie, and Y . Qu, “Relationship prompt learning is enough for open-vocabulary semantic segmentation,”Advances in Neural Infor- mation Processing Systems, vol. 37, pp. 74 298–74 324, 2024

    2024

  7. [7]

    More-brain: Routed mixture of experts for interpretable and generalizable cross-subject fmri visual decoding,

    Y . Wei, Y . Zhang, X. Xiao, T. Wang, X. Wang, and V . D. Calhoun, “More-brain: Routed mixture of experts for interpretable and generalizable cross-subject fmri visual decoding,”arXiv preprint arXiv:2505.15946, 2025

    work page arXiv 2025

  8. [8]

    Ddfp: Data-dependent frequency prompt for source free domain adaptation of medical image segmentation,

    S. Yin, S. Liu, and M. Wang, “Ddfp: Data-dependent frequency prompt for source free domain adaptation of medical image segmentation,” Knowledge-Based Systems, p. 113651, 2025

    2025

  9. [9]

    Pre- train, prompt, and predict: A systematic survey of prompting methods in natural language processing,

    P. Liu, W. Yuan, J. Fu, Z. Jiang, H. Hayashi, and G. Neubig, “Pre- train, prompt, and predict: A systematic survey of prompting methods in natural language processing,”ACM computing surveys, vol. 55, no. 9, pp. 1–35, 2023

    2023

  10. [10]

    Few-shot parameter-efficient fine-tuning is better and cheaper than in-context learning,

    H. Liu, D. Tam, M. Muqeeth, J. Mohta, T. Huang, M. Bansal, and C. A. Raffel, “Few-shot parameter-efficient fine-tuning is better and cheaper than in-context learning,”Advances in Neural Information Processing Systems, vol. 35, pp. 1950–1965, 2022

    1950

  11. [11]

    Autovp: An automated visual prompting framework and benchmark,

    H.-A. Tsao, L. Hsiung, P.-Y . Chen, S. Liu, and T.-Y . Ho, “Autovp: An automated visual prompting framework and benchmark,”arXiv preprint arXiv:2310.08381, 2023

    work page arXiv 2023

  12. [12]

    Lor-vp: Low-rank visual prompting for efficient vision model adaptation,

    C. Jin, Y . Li, M. Zhao, S. Zhao, Z. Wang, X. He, L. Han, T. Che, and D. N. Metaxas, “Lor-vp: Low-rank visual prompting for efficient vision model adaptation,”arXiv preprint arXiv:2502.00896, 2025

    work page arXiv 2025

  13. [13]

    Understanding and improving visual prompting: A label-mapping perspective,

    A. Chen, Y . Yao, P.-Y . Chen, Y . Zhang, and S. Liu, “Understanding and improving visual prompting: A label-mapping perspective,” in Proceedings of the IEEE/CVF Conference on Computer Vision and Pattern Recognition, 2023, pp. 19 133–19 143

    2023

  14. [14]

    Visual prompt tuning,

    M. Jia, L. Tang, B.-C. Chen, C. Cardie, S. Belongie, B. Hariharan, and S.-N. Lim, “Visual prompt tuning,” inEuropean conference on computer vision. Springer, 2022, pp. 709–727

    2022

  15. [15]

    Insvp: Efficient instance visual prompting from image itself,

    Z. Liu, Y . Peng, and J. Zhou, “Insvp: Efficient instance visual prompting from image itself,” inProceedings of the 32nd ACM International Conference on Multimedia, 2024, pp. 6443–6452

    2024

  16. [16]

    Da-vpt: Semantic-guided visual prompt tuning for vision transformers,

    L. Ren, C. Chen, L. Wang, and K. Hua, “Da-vpt: Semantic-guided visual prompt tuning for vision transformers,” inProceedings of the Computer Vision and Pattern Recognition Conference, 2025, pp. 4353–4363

    2025

  17. [17]

    Top- kast: Top-k always sparse training,

    S. Jayakumar, R. Pascanu, J. Rae, S. Osindero, and E. Elsen, “Top- kast: Top-k always sparse training,”Advances in Neural Information Processing Systems, vol. 33, pp. 20 744–20 754, 2020

    2020

  18. [18]

    E2vpt: An effective and efficient approach for visual prompt tuning. in 2023 ieee,

    C. Han, Q. Wang, Y . Cui, Z. Cao, W. Wang, S. Qi, and D. Liu, “E2vpt: An effective and efficient approach for visual prompt tuning. in 2023 ieee,” inCVF International Conference on Computer Vision (ICCV), 2023, pp. 17 445–17 456

    2023

  19. [19]

    Dropout: a simple way to prevent neural networks from over- fitting,

    N. Srivastava, G. Hinton, A. Krizhevsky, I. Sutskever, and R. Salakhut- dinov, “Dropout: a simple way to prevent neural networks from over- fitting,”The journal of machine learning research, vol. 15, no. 1, pp. 1929–1958, 2014

    1929

  20. [20]

    Network anomaly detection using threshold-based sparse,

    M. T. Tun, D. E. Nyaung, and M. P. Phyu, “Network anomaly detection using threshold-based sparse,” inProceedings of the 11th International conference on advances in information technology, 2020, pp. 1–8

    2020

  21. [21]

    Exploring sparse visual prompt for domain adaptive dense prediction,

    S. Yang, J. Wu, J. Liu, X. Li, Q. Zhang, M. Pan, Y . Gan, Z. Chen, and S. Zhang, “Exploring sparse visual prompt for domain adaptive dense prediction,” inProceedings of the AAAI Conference on Artificial Intelligence, vol. 38, no. 15, 2024, pp. 16 334–16 342

    2024

  22. [22]

    Sharing leaky-integrate- and-fire neurons for memory-efficient spiking neural networks,

    Y . Kim, Y . Li, A. Moitra, R. Yin, and P. Panda, “Sharing leaky-integrate- and-fire neurons for memory-efficient spiking neural networks,”Fron- tiers in Neuroscience, vol. 17, p. 1230002, 2023

    2023

  23. [23]

    Direct training high-performance deep spiking neural networks: a review of theories and methods,

    C. Zhou, H. Zhang, L. Yu, Y . Ye, Z. Zhou, L. Huang, Z. Ma, X. Fan, H. Zhou, and Y . Tian, “Direct training high-performance deep spiking neural networks: a review of theories and methods,”Frontiers in Neuroscience, vol. 18, p. 1383844, 2024

    2024

  24. [24]

    When prompting meets spiking: Graph sparse prompting via spiking graph prompt learning,

    B. Jiang, W. Zhao, B. Wang, and J. Tang, “When prompting meets spiking: Graph sparse prompting via spiking graph prompt learning,” arXiv preprint arXiv:2601.02662, 2026

    work page arXiv 2026

  25. [25]

    Networks of spiking neurons: the third generation of neural network models,

    W. Maass, “Networks of spiking neurons: the third generation of neural network models,”Neural networks, vol. 10, no. 9, pp. 1659–1671, 1997

    1997

  26. [26]

    Pts-snn: A prompt-tuned temporal shift spiking neural networks for efficient speech emotion recognition,

    X. Su, H. Wang, and Q. Zhang, “Pts-snn: A prompt-tuned temporal shift spiking neural networks for efficient speech emotion recognition,”arXiv preprint arXiv:2602.08240, 2026

    work page arXiv 2026

  27. [27]

    Snnptrack: Spiking neural network based prompt for high-accuracy rgbe tracking,

    Y . Ji, Q. Zhao, Y . Liang, and J. Wu, “Snnptrack: Spiking neural network based prompt for high-accuracy rgbe tracking,” inICASSP 2025-2025 IEEE International Conference on Acoustics, Speech and Signal Processing (ICASSP), 2025, pp. 1–5

    2025

  28. [28]

    Unleashing the power of visual prompting at the pixel level,

    J. Wu, X. Li, C. Wei, H. Wang, A. Yuille, Y . Zhou, and C. Xie, “Unleashing the power of visual prompting at the pixel level,”arXiv preprint arXiv:2212.10556, 2022

    work page arXiv 2022

  29. [29]

    Bayesian-guided label mapping for visual reprogramming,

    C. Cai, Z. Ye, L. Feng, J. Qi, and F. Liu, “Bayesian-guided label mapping for visual reprogramming,” inThe Thirty-eighth Annual Conference on Neural Information Processing Systems, 2024

    2024

  30. [30]

    Attribute-based visual reprogramming for vision-language models,

    C. Cai, Z. Ye, L. Feng, and J. Qi, “Attribute-based visual reprogramming for vision-language models,”arXiv preprint arXiv:2501.13982, 2025

    work page arXiv 2025

  31. [31]

    Ot-vp: Optimal transport-guided visual prompting for test-time adaptation,

    Y . Zhang, A. Mehra, and J. Hamm, “Ot-vp: Optimal transport-guided visual prompting for test-time adaptation,” in2025 IEEE/CVF Winter Conference on Applications of Computer Vision (WACV). IEEE, 2025, pp. 1122–1132

    2025

  32. [32]

    Blackvip: Black-box visual prompting for robust transfer learning,

    C. Oh, H. Hwang, H.-y. Lee, Y . Lim, G. Jung, J. Jung, H. Choi, and K. Song, “Blackvip: Black-box visual prompting for robust transfer learning,” inProceedings of the IEEE/CVF Conference on Computer Vision and Pattern Recognition, 2023, pp. 24 224–24 235

    2023

  33. [33]

    Transfer learning without know- ing: Reprogramming black-box machine learning models with scarce data and limited resources,

    Y .-Y . Tsai, P.-Y . Chen, and T.-Y . Ho, “Transfer learning without know- ing: Reprogramming black-box machine learning models with scarce data and limited resources,” inInternational Conference on Machine Learning. PMLR, 2020, pp. 9614–9624

    2020

  34. [34]

    Prompt tuning for unified multimodal pretrained models,

    H. Yang, J. Lin, A. Yang, P. Wang, and C. Zhou, “Prompt tuning for unified multimodal pretrained models,” inFindings of the Association for Computational Linguistics: ACL 2023, 2023, pp. 402–416

    2023

  35. [35]

    Spiking transformer: Introducing accurate addition-only spiking self-attention for transformer,

    Y . Guo, X. Liu, Y . Chen, W. Peng, Y . Zhang, and Z. Ma, “Spiking transformer: Introducing accurate addition-only spiking self-attention for transformer,” inProceedings of the Computer Vision and Pattern Recognition Conference, 2025, pp. 24 398–24 408

    2025

  36. [36]

    Spike- driven transformer,

    M. Yao, J. Hu, Z. Zhou, L. Yuan, Y . Tian, B. Xu, and G. Li, “Spike- driven transformer,”Advances in neural information processing systems, vol. 36, pp. 64 043–64 058, 2023

    2023

  37. [37]

    Event-driven learning for spiking neural networks,

    W. Wei, M. Zhang, J. Zhang, A. Belatreche, J. Wu, Z. Xu, X. Qiu, H. Chen, Y . Yang, and H. Li, “Event-driven learning for spiking neural networks,”arXiv preprint arXiv:2403.00270, 2024

    work page arXiv 2024

  38. [38]

    Spiking deep convolutional neural networks for energy-efficient object recognition,

    Y . Cao, Y . Chen, and D. Khosla, “Spiking deep convolutional neural networks for energy-efficient object recognition,”International Journal of Computer Vision, vol. 113, no. 1, pp. 54–66, 2015

    2015

  39. [39]

    Spiking de- noising diffusion probabilistic models,

    J. Cao, Z. Wang, H. Guo, H. Cheng, Q. Zhang, and R. Xu, “Spiking de- noising diffusion probabilistic models,” inProceedings of the IEEE/CVF winter conference on applications of computer vision, 2024, pp. 4912– 4921

    2024

  40. [40]

    Integer-valued training and spike-driven inference spiking neural network for high-performance and energy-efficient object detection,

    X. Luo, M. Yao, Y . Chou, B. Xu, and G. Li, “Integer-valued training and spike-driven inference spiking neural network for high-performance and energy-efficient object detection,” inEuropean Conference on Computer Vision. Springer, 2024, pp. 253–272

    2024

  41. [41]

    Recdis-snn: Rectifying membrane potential distribution for directly training spiking neural networks,

    Y . Guo, X. Tong, Y . Chen, L. Zhang, X. Liu, Z. Ma, and X. Huang, “Recdis-snn: Rectifying membrane potential distribution for directly training spiking neural networks,” inProceedings of the IEEE/CVF conference on computer vision and pattern recognition, 2022, pp. 326– 335

    2022

  42. [42]

    Spiking-yolo: spiking neural network for energy-efficient object detection,

    S. Kim, S. Park, B. Na, and S. Yoon, “Spiking-yolo: spiking neural network for energy-efficient object detection,” inProceedings of the AAAI conference on artificial intelligence, vol. 34, no. 07, 2020, pp. 11 270–11 277

    2020

  43. [43]

    Spiking pointnet: Spiking neural networks for point clouds,

    D. Ren, Z. Ma, Y . Chen, W. Peng, X. Liu, Y . Zhang, and Y . Guo, “Spiking pointnet: Spiking neural networks for point clouds,”Advances in Neural Information Processing Systems, vol. 36, pp. 41 797–41 808, 2023

    2023

  44. [44]

    Temporal-guided spiking neural networks for event-based human action recognition,

    S. Yang, S. Lu, S. Wang, M. H. Er, Z. Zheng, and A. C. Kot, “Temporal-guided spiking neural networks for event-based human action recognition,”arXiv preprint arXiv:2503.17132, 2025

    work page arXiv 2025

  45. [45]

    Dynamic weight adaptation in spiking neural networks inspired by JOURNAL OF LATEX CLASS FILES, VOL. 14, NO. 8, AUGUST 2021 10 biological homeostasis,

    Y . Zhou, B. Dong, C. Li, Y . Wang, X. Yin, Y . Wang, and X. Yang, “Dynamic weight adaptation in spiking neural networks inspired by JOURNAL OF LATEX CLASS FILES, VOL. 14, NO. 8, AUGUST 2021 10 biological homeostasis,” inProceedings of the AAAI Conference on Artificial Intelligence, vol. 40, no. 34, 2026, pp. 29 089–29 097

    2021

  46. [46]

    Lora: Low-rank adaptation of large language models

    E. J. Hu, Y . Shen, P. Wallis, Z. Allen-Zhu, Y . Li, S. Wang, L. Wang, W. Chenet al., “Lora: Low-rank adaptation of large language models.” Iclr, vol. 1, no. 2, p. 3, 2022

    2022

  47. [47]

    Tiny imagenet visual recognition challenge,

    Y . Le and X. Yang, “Tiny imagenet visual recognition challenge,”CS 231N, vol. 7, no. 7, p. 3, 2015

    2015

  48. [48]

    Learning multiple layers of features from tiny images.(2009),

    A. Krizhevsky, G. Hintonet al., “Learning multiple layers of features from tiny images.(2009),” 2009

    2009

  49. [49]

    Eurosat: A novel dataset and deep learning benchmark for land use and land cover classi- fication,

    P. Helber, B. Bischke, A. Dengel, and D. Borth, “Eurosat: A novel dataset and deep learning benchmark for land use and land cover classi- fication,”IEEE Journal of Selected Topics in Applied Earth Observations and Remote Sensing, vol. 12, no. 7, pp. 2217–2226, 2019

    2019

  50. [50]

    Cats and dogs,

    O. M. Parkhi, A. Vedaldi, A. Zisserman, and C. Jawahar, “Cats and dogs,” in2012 IEEE conference on computer vision and pattern recog- nition, 2012, pp. 3498–3505

    2012

  51. [51]

    Food-101–mining dis- criminative components with random forests,

    L. Bossard, M. Guillaumin, and L. Van Gool, “Food-101–mining dis- criminative components with random forests,” inProceedings of the IEEE Conference on Computer vision ECCV 2014, 2014, pp. 446–461

    2014

  52. [52]

    Describing textures in the wild,

    M. Cimpoi, S. Maji, I. Kokkinos, S. Mohamed, and A. Vedaldi, “Describing textures in the wild,” inProceedings of the IEEE Conference on Computer Vision and Pattern Recognition, 2014, pp. 3606–3613

    2014

  53. [53]

    Automated flower classification over a large number of classes,

    M.-E. Nilsback and A. Zisserman, “Automated flower classification over a large number of classes,” in2008 Sixth Indian conference on computer vision, graphics & image processing, 2008, pp. 722–729

    2008

  54. [54]

    Imagenet: A large-scale hierarchical image database,

    J. Deng, W. Dong, R. Socher, L.-J. Li, K. Li, and L. Fei-Fei, “Imagenet: A large-scale hierarchical image database,” in2009 IEEE conference on computer vision and pattern recognition, 2009, pp. 248–255

    2009

  55. [55]

    Do imagenet clas- sifiers generalize to imagenet?

    B. Recht, R. Roelofs, L. Schmidt, and V . Shankar, “Do imagenet clas- sifiers generalize to imagenet?” inInternational conference on machine learning, 2019, pp. 5389–5400

    2019

  56. [56]

    Learning robust global representations by penalizing local predictive power,

    H. Wang, S. Ge, Z. Lipton, and E. P. Xing, “Learning robust global representations by penalizing local predictive power,”Advances in Neural Information Processing Systems, vol. 32, 2019

    2019

  57. [57]

    Natural adversarial examples,

    D. Hendrycks, K. Zhao, S. Basart, J. Steinhardt, and D. Song, “Natural adversarial examples,” inProceedings of the IEEE Conference on Computer Vision and Pattern Recognition, 2021, pp. 15 262–15 271

    2021

  58. [58]

    The many faces of robustness: A critical analysis of out-of-distribution generalization,

    D. Hendrycks, S. Basart, N. Mu, S. Kadavath, F. Wang, E. Dorundo, R. Desai, T. Zhu, S. Parajuli, M. Guoet al., “The many faces of robustness: A critical analysis of out-of-distribution generalization,” in Proceedings of the IEEE Conference on Computer Vision and Pattern Recognition, 2021, pp. 8340–8349

    2021

  59. [59]

    Deep residual learning for image recognition,

    K. He, X. Zhang, S. Ren, and J. Sun, “Deep residual learning for image recognition,” inProceedings of the IEEE conference on computer vision and pattern recognition, 2016, pp. 770–778

    2016

  60. [60]

    Learning transferable visual models from natural language supervision,

    A. Radford, J. W. Kim, C. Hallacy, A. Ramesh, G. Goh, S. Agarwal, G. Sastry, A. Askell, P. Mishkin, J. Clarket al., “Learning transferable visual models from natural language supervision,” inInternational conference on machine learning. PmLR, 2021, pp. 8748–8763

    2021

  61. [61]

    Un- leashing the power of visual prompting at the pixel level,

    J. Wu, X. Li, C. Wei, H. Wang, A. Yuille, Y . Zhou, and C. Xie, “Un- leashing the power of visual prompting at the pixel level,”Transactions on Machine Learning Research, 2024

    2024

  62. [62]

    Large-scale machine learning with stochastic gradient de- scent,

    L. Bottou, “Large-scale machine learning with stochastic gradient de- scent,” inProceedings of COMPSTAT’2010: 19th International Con- ference on Computational StatisticsParis France, August 22-27, 2010 Keynote, Invited and Contributed Papers. Springer, 2010, pp. 177– 186

    2010

  63. [63]

    Maple: Multi-modal prompt learning,

    M. U. Khattak, H. Rasheed, M. Maaz, S. Khan, and F. S. Khan, “Maple: Multi-modal prompt learning,” inProceedings of the IEEE/CVF conference on computer vision and pattern recognition, 2023, pp. 19 113–19 122

    2023